not my answer, but see:

http://astronautix.com/Mambo/

I don't know how the new AF engine would work, but using the SSME for TLI strikes me as being okay. SSME is the US's most advanced engine, and anything better could probably best be built on the same production lines.

But that's just my opinion and a quick guess at that - any real analysis would have to be much deeper and much more thorough.

The big question for me right now is what is the maximum storage time for LOX? Hypergolics are limited to six months storage because of gasket deterioration, (hypergolics on comsats must last longer, but I think most of these are monopropellant), but what is the current state of the art maximum storage time for LOX?

WE could develop the CEV and TLI stages while operating the shuttle though to 2011 and completing the ISS, fulfilling our international obligataions, all without breaking the budget. SDLV development could be put off until 2012, after the shuttle is shut down. The new AF engines could then be used for this, if they are viable.

That's the best I can come up with.